GNU bug report logs - #15045
Point jumps inappropriately around time of Semantic lexing

Previous Next

Full log

Message #212 received at 15045 <at> debbugs.gnu.org (full text, mbox):

From: Barry OReilly <gundaetiapo <at> gmail.com>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 15045 <at> debbugs.gnu.org,
 David Engster <deng <at> randomsample.de>, Eric Ludlam <eric <at> siege-engine.com>
Subject: Re: bug#15045: Point jumps inappropriately around time of Semantic
 lexing
Date: Fri, 25 Oct 2013 15:15:48 -0400

[Message part 1 (text/plain, inline)]

In terms of solving the issue of how Semantic calls
accept-process-output, here are two proposals:

1: Per David Engster:

> If you look in the `define-lex' macro, you see that it calls
> `semantic-throw-on-input' after each identified token. The problem is
> that it does not restore the cursor position before that, so I guess the
> fix is simply to change this call to
>
>        (save-excursion
>          (goto-char starting-position)
>          (semantic-throw-on-input 'lex))

Here's a patch to implement the basic idea. My debugging indicates
accept-process-output can be called when lexing other buffers, so I
saved off the marker rather than just position.

That said, I have never personally witnessed point unexpectedly
redisplayed in another buffer -- only within the same buffer. That
could be a consequence of how it interacts with the
jit-lock-deferred-fontify idle timer. (I have confirmed with my debug
statements that Semantic is the outer timer and
jit-lock-deferred-fontify is the inner timer which redisplays.)

Please let me know if this fix is suitable.

diff --git a/lisp/cedet/semantic/fw.el b/lisp/cedet/semantic/fw.el
index 825cdc9..869d183 100644
--- a/lisp/cedet/semantic/fw.el
+++ b/lisp/cedet/semantic/fw.el
@@ -369,6 +369,8 @@ later installation should be done in MODE hook."
 ;;
 (defvar semantic-current-input-throw-symbol nil
   "The current throw symbol for `semantic-exit-on-input'.")
+(defvar semantic--on-input-start-marker nil
+  "The marker when starting a semantic-exit-on-input form.")

 (defmacro semantic-exit-on-input (symbol &rest forms)
   "Using SYMBOL as an argument to `throw', execute FORMS.
@@ -376,7 +378,8 @@ If FORMS includes a call to `semantic-throw-on-input',
then
 if a user presses any key during execution, this form macro
 will exit with the value passed to `semantic-throw-on-input'.
 If FORMS completes, then the return value is the same as `progn'."
-  `(let ((semantic-current-input-throw-symbol ,symbol))
+  `(let ((semantic-current-input-throw-symbol ,symbol)
+         (semantic--on-input-start-marker (point-marker)))
      (catch ,symbol
        ,@forms)))
 (put 'semantic-exit-on-input 'lisp-indent-function 1)
@@ -387,7 +390,16 @@ FROM is an indication of where this function is called
from as a value
 to pass to `throw'.  It is recommended to use the name of the function
 calling this one."
   `(when (and semantic-current-input-throw-symbol
-              (or (input-pending-p) (accept-process-output)))
+              (or (input-pending-p)
+                  (save-excursion
+                    ;; Timers might run during accept-process-output.
+                    ;; If they redisplay, point must be where the user
+                    ;; expects. (Bug#15045)
+                    (set-buffer (marker-buffer
+                                 semantic--on-input-start-marker))
+                    (goto-char (marker-position
+                                semantic--on-input-start-marker))
+                    (accept-process-output))))
      (throw semantic-current-input-throw-symbol ,from)))

 ^L

2: Extend accept-process-output to allow suppressing timer_check.
Actually, there's an input arg already for that, but only applicable
if the PROCESS arg is non nil.

[Message part 2 (text/html, inline)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.